There are several potential mechanisms leading to development of malignancies. It is also not completely clear at present whether cancer originates in mutated already differentiated somatic cells or occurs from the mutated stem/progenitor cells [1–4]. In several types of tumors there have been identified cells that express primitive phenotype and postulated to act as cancer stem cells (CSCs) [5–10]. The potential origin of these CSCs is intriguing and their potential developmental relationship to normal non-malignant stem cells remains unclear. In particular special attention deserve identified in adult tissues population of early development stem cells [11–16], because such cells could be origin of some primitive tumors.

Nevertheless, the concept that adult tissues contain development early cells with embryonic features that can lead to tumors is, surprisingly, not so novel. During the 19th and early 20th centuries, several scientists proposed that cancer may develop in populations of cells that are left in a dormant state in developing organs during embryogenesis. This assumption lend initially basis to so-called “embryonic rest hypothesis of cancer origin” that was proposed by Recamier in 1829 [17] and Remak in 1854 [18] and most extensively elaborated in its final form by Virchow in 1855 [19], Durante in 1874 [20], and Cohnheim in 1875 [21]. However, the putative cells responsible for this effect remained at that time somehow elusive and neither clearly identified nor purified from adult tissues.

According to “embryonic rest hypothesis of cancer origin” adult tissues contain embryonic remnants that normally lie dormant but can be “activated” to become malignant. In agreement what has been postulated by Virchow and Cohnheim in the middle of XIX century, subsequently Wright in 1910 [22] proposed a germinal cell origin of Wilm's tumor (nephroblastoma) and Beard in 1911 [23] postulated that tumors arise from displaced and activated trophoblasts or displaced germline cells. Although, the term “stem cell” was not widely used at that time in scientific language, it is reasonable to assume that they believed that such cells would be endowed with properties similar to stem cell properties as we understand them today.

In this review, we present augmenting evidence regarding the existence of a development primitive population of so-called very small embryonic/epiblast-like stem cells (VSELs) deposited in early-developing tissues during organogenesis, and the developmental relationship of VSELs to the epiblast/germ lineage. We will also present some crucial epigenetic mechanisms based on silencing of imprinted genes involved in insulin/insulin like growth factor signaling (IIS) that control VSEL proliferation and prevent “unleashed” expansion of these cells. We will also discuss other potential mechanisms such as involvement of VSELs in creation of aneuploid cells and in supporting stromalization and vascularization of growing tumors.

The hypothesis presented in this review, however, needs further experimental support and in the present form provides a basis for future experimentations. Our goal is to encourage other colleagues to consider the possibility that VSELs and other types of developmentally primitive stem cells deposited in adult tissues, such as for example similar to VSELs or even overlapping populations of multipotent adult stem cells (MASCs), multipotent adult progenitor cells (MAPCs) or marrow-isolated adult multilineage-inducible cells (MIAMI), could be involved in cancer development.

It is widely accepted that organ/tissue regeneration and cancer development are likely to be closely related processes [24]. In support of this notion carcinogenesis is very often a response to chronic irritation, inflammation, and tissue injury or damage, potentially developing through misappropriation of homeostatic mechanisms that govern normal tissue repair and stem cell self-renewal [25]. Indeed, cancer incidence increases in many cases associated with chronic tissue or organ injury. These observations strongly support the involvement of tissue repair-related mechanisms in the development of some malignancies. This also suggests a potential role for normal stem cells in origination of cancer stem cells (CSCs) [26].

If some malignancies originate in stem cell compartment, it is unclear how developmentally primitive these cells of origin are. It is well known that the stem cell compartment in adult tissues displays a developmental hierarchy. Besides already tissue-committed stem cells (TCSCs) such as for example hematopoietic stem cells (HSCs), muscle stem cells, or epidermal stem cells, which give rise to the cells present in a given tissue, evidence has accumulated that more primitive stem cells with a broader ability to differentiate also exist [11–16]. Based on observations that adult tissues contain development early stem cells (e.g., VSELs, MASCs, MAPCs and MIAMI), these cells could be the missing link that reconciles the embryonic rest hypothesis of cancer origin postulated by Virchow [19] with current theories postulating cancer as a stem cell disorder.

The hypothesis that CSCs develop in the normal stem/progenitor cell compartment is based on the assumption that self-renewing normal stem cells residing in organs (e.g., bronchial or stomach epithelium), and not mature differentiated somatic cells such as those lining the bronchial or stomach mucosa [27, 28], may not only acquire but more important accumulate mutations during life. While differentiated mucosa cells if they acquire mutation are shed from epithelium and are eliminated, mutations if they occur in stem cells would be maintained in the stem cell compartment and passed on daughter stem cells. Moreover, self-renewing stem cells may then acquire additional mutations and epigenetic changes so that the genome is destabilized; at a certain point, CSC will emerge and an uncontrolled neoplastic proliferation may be initiated.

Indeed, as mentioned above recent evidence suggests that malignancy arises from the accumulation of mutations and maturation arrest of normal stem/progenitor cells rather than by the dedifferentiation of already differentiated cells [2, 3, 27, 29, 30]. Thus, normal stem cells may acquire mutations and give rise to CSCs, which are subsequently responsible for i) initial tumor growth, ii) tumor re-growth after unsuccessful radio-chemotherapy, and iii) establishing distant metastases. In support of this, recent research has provided direct evidence that several neoplasms (e.g., brain tumors, prostate cancer, melanomas, colon, and lung cancer) may in fact originate in the normal stem cells [3, 27, 31, 32]. Specifically, rare populations of cells with stem cell properties [(e.g., expression of cell surface stem cell antigens (CD133), expression of some pluripotent stem cell markers (Oct-4) or their location in so called side population of cells after Hoe33342 staining)] have been identified within growing malignancies. Moreover such cells, after inoculation into immunodeficient mice, are able to give rise to tumors that morphologically resemble those tumors from which they were initially purified [33]. However, it is still difficult to demonstrate such a phenomenon at the single-cell level, and usually a minimum number of cells has to be inoculated to give rise to the tumor, suggesting the importance of crosstalk between the primitive cells that initiate these malignancies.

We postulate that from the developmental point of view, the germ lineage is an origin and “scaffold” of stem cells in the adult body. It is well known that germ line is somehow immortal, as it passes genomic and mitochondrial DNA to the next generation [34, 35]. From biological point of view germline during embryogenesis creates mortal soma and body that are essential for the germline to fulfill its reproductive mission.

Fertilized oocyte or zygote is the most primitive cell at the top of the germline hierarchy. The germline potential is subsequently maintained during the first steps of development in the blastomeres of the morula and in the inner cell mass (ICM) of the developing blastocyst. At the level of the blastocyst, however, a portion of the cells surrounding the blastula segregates from the germline lineage and differentiates into the trophoblast that will develop into the placenta. At the same time, the rest of the cells in the ICM retain their germline character, and after implantation of the blastocyst in the uterus, the ICM cells give rise to the epiblast. Epiblast contains pluripotent stem cells (PSCs) that also retain germline potential. While a group of cells in proximal epiblast specifies into primordial germ cells (PGCs) which migrate to the genital ridges and give rise to the precursors of gametes [36, 37], the other epiblast PSCs become specified into stem cells for meso-, ecto-, and endoderm, which later give rise to TCSCs for all the developing organs and tissues in the embryo proper.

Recently identified, VSELs express several markers characteristic of the epiblast and germ line [38, 39]. Accordingly, the TEM studies revealed VSELs to be small cells with a high nuclear/cytoplasmic ratio, euchromatin, and few mitochondria [12]. VSELs isolated under steady-state conditions from highly express, at the mRNA and protein levels, genes involved in specification of the epiblast (e.g., Stella, Fragilis, Blimp1) in addition to genes involved in PGCs specification, such as Dppa2, Dppa4, and Mvh, which characterize late-migratory PGCs [38]. The expression of some of these genes has been confirmed at the protein level and at the promoter level to confirm chromatin structure characteristic of actively transcribed genes.

More importantly, we also observed that VSELs and HSCs express mRNA for several pituitary and gonadal hormone receptors as well as highly express Sall4, an early-development marker of germ cells [40, 41]. Finally, in direct in vitro and in vivo experiments, we confirmed that the quiescent population of BM-residing VSELs responds to stimulation by androgens (danazol) and pituitary gonadotropins – follicle-stimulating hormone (FSH) and luteinizing hormone (LH). In particular, we found that 10-day administration of androgens or FSH and LH directly stimulated expansion (∼2–3x) of VSELs and HSPCs in BM and enhanced BrdU incorporation (K Mierzejewska, manuscript in preparation).

Thus our data support the VSELs that express several unique epiblast and PGCs markers, are the most primitive population of stem cells in BM and represent cells that are related to germ line. We envision as mentioned above that VSELs, could be the missing link that reconciles the embryonic rest hypothesis of cancer origin postulated by Virchow [19] with current theories postulating cancer as a stem cell disorder.

Evidence accumulates that supports the embryonic rest hypothesis of cancer development and the potential involvement of epiblast/germline cells in this process. Firstly, it is the existence of “classical” germline tumors such as seminomas, ovarian tumors, yolk sac tumor, mediastinal or brain germ cell tumors, teratomas, and teratocarcinomas. Secondly, in several types of cancer cells types (e.g., gastric, lung, liver, and bladder carcinomas as well as melanomas, medulloblastomas, pediatric sarcomas, and germinal tumors), expression of cancer testis (C/T) antigens has been observed (∼40 identified so far), which are encoded by genes that are normally expressed only in the human germline [42, 43]. Thirdly, many malignancies express some typical embryonic cells markers, such as chorionic gonadotropin (hCG) and/or carcinoembryonic antigen (CEA), which as it is well known are employed in cancer diagnostics [44]. Finally, several solid tumors express the embryonic/germline transcription factor Oct-4, a marker of embryonic, epiblast and germline PSCs, that as reported is expressed in various tumor types (e.g., gastric, lung, bladder, and oral mucosa carcinomas and in germinal tumors) [45, 46].

Thus, based on this some malignancies could develop from very primitive cells retained in adult tissues. There is also a parallel possibility that all these abovementioned changes emerge in cancer cells due to epigenetic reprograming of somatic cells during malignant transformation so that they acquire epiblast/germline markers.

We hypothesize that primitive types of solid tumors may develop from the most primitive from a developmental point of view cells and could well be VSELs. To support this VSELs as mentioned above express not only several markers characteristic of the epiblast/germline but also highly express mRNAs for several C/T antigens [47]. However, VSELs residing in adult tissues are in highly quiescent state. We have reported that these cells like migrating PGCs undergo epigenetic changes, due to modification of genomic imprinting of some early-development parentally imprinted gene loci, including Igf2-H19 and KCNQ1/p57^Kip2, which results in their resistance to insulin/insulin like growth factors signaling (IIS) and upregulation of the cyclin-dependent kinase inhibitor p57^KIP2 [48]. In addition to changes in expression of imprinted genes, VSELs express several miRNAs that attenuate IIS in these cells (mir681, mir470, mir669b) as well as upregulate expression of p57^KIP2 (mir25.1, mir19b, mir92) (not published). All these epigenetic changes are responsible for quiescent state of VSELs.

Figure 1 depicts several possibilities how VSELs could participate in tumor development and all these possibilities will be discussed below. However, at beginning we have to admit that better direct evidence is still needed to show that VSELs or pluripotent/multipotent populations of stem cells related to them residing in adult tissues are truly a source of malignancies.

Figure 1. Potential involvement of VSELs in carcinogenesis. VSELs may give rise to cancer (stem) cells (red irregular cell) if their genomic imprinting is not erased (teratomas/teratocarcinomas) or lost (e.g., rhabdomyosarcoma, neuroblastoma, nephroblastoma). We envision that this particular scenario plays a significant role in pathogenesis of pediatric tumors. VSELs as other cells could also be a direct target for other type of mutations. VSELs may fuse due to the chronic inflammation/fusogenic agent within a normal somatic cell (light grey box with dark nucleus) and give rise to a heterokaryon (blue cells with dark nucleus). The tetraploid heterokaryon undergoes subsequent selection and gives rise to the aneuploid cancer (stem) cell. We envision that this scenario plays a more important role in pathogenesis of tumors developing in adult patients. VSELs may indirectly contribute to tumorogenesis by providing vasculature and stroma cells for growing tumor tissue. This may occur in tumors developing both in young and older patients.

We reported that one of the important potential connections between tumorogenesis and VSELs is epigenetic modification at the insulin-like growth factor-2 (Igf2)-H19 locus [48, 49]. Erasure of parental imprinting (EOI) at this locus keeps these cells quiescent in adult tissues [48] and for example, teratomas and teratocarcinomas may originate in VSELs residing in adult tissues as a result of persistent somatic imprinting due to lack of erasure of imprinting, in particular at the Igf2-H19 locus. On other hand hypermethylation of regulatory elements for Igf2-H19 locus known as loss of imprinting (LOI) occurs in some malignancies such as for example rhabdomyosarcoma, neuroblastoma, Ewing's sarcoma, and Wilm's tumor (nephroblastoma). Moreover, it is possible that some typical germinal tumors, including germinomas, seminomas, teratomas, dermoid cyst, and hydatidiform mole, most likely develop in mutated cells related to VSELs left along the PGC migratory route from the primitive streak to the genital ridges. In addition as mentioned above, the fact that VSELs respond to FSH, whose level is upregulated with advanced age, may somehow render these cells sensitive to malignant transformation and may explain age-related increase in malignancies.

In the past years several groups have proposed a concept that cancer is a chromosomal rather than a genetic disease. Accordingly, carcinogenesis could be initiated by aneuploidies induced by cancer genes, which unbalance genes expression and lead over time to the selection of aneuploid clones of transformed cells [50]. In this context, VSELs, as highly fusogenic cells, could be potential fusion partners for somatic cells. VSELs could provide several transcripts characteristic for early developmental cells (e.g., Oct-4, Klf-4), while somatic cells could supply chromosomes that show proper genomic imprinting and tissue specification. The formation of such heterokaryons could be a first step in the subsequent selection of anueploid “immortal” cells. In most of the cases such fusion-created heterokaryons will not survive and are eliminated, but it may happen that by deleting some additional chromosomes they will reach an aneuploidy state that will favor survival and clonal expansion of anueploid cell.

For example, in stomach or lung cancer may originate if circulating in peripheral blood or tissue- residing VSELs fuse with differentiated cells in Helicobacter pylori-infected stomach or tobacco smoke-damaged lung respectively [28, 51].

VSELs have been demonstrated to give rise both to the precursor cells of mesenchymal lineage as well as to endothelial progenitor cells [52]. Thus, if expanding tumor secrete chemoattractants for VSELs, population of these cells that circulate in could be chemoattracted by the expanding cancer tissue and potentially involved in tumor expansion and growth by supplying vessels and stroma. Chemotactic factors that are identified and are involved in these phenomena include stromal derived factor-1 (SDF-1), hepatocyte growth factor/scatter factor (HGF/SF) as well as some bioactive phosphosphingolipids such as sphingosine-1-phosphate (S1P) or ceramide-1-phosphate (C1P) [12, 53, 54]. All these factors have been demonstrated to be not only chemoattractants for VSELs but as demonstrated are also secreted by hypoxic tissues [55–57]. In fact recent evidence indicates that VSELs circulate in PB both in cancer patients as well as in experimental models with tumors inoculated into mice [47, 58].

We can assume that VSELs circulating in PB are chemoattracted to tumor tissues, being biased by chemotactic signals and wrongly recognize the expanding tumor as regenerating tissue that needs stroma and vessel support. Thus, VSELs are incorporated into the tumor microenvironment “in good faith” and unfortunately provide vessels and stroma to the wrong pathological tissue [47].

There are several indications that some cancers originate in cells closely related to the eraly development epiblast or germline cells. Our team recently identified a unique population of VSELs that express several epiblast/germline markers, implying that these rare cells could in fact be the source of several malignancies. Accordingly, it is very likely that the VSELs that express several germline markers (e.g., Oct-4, MvH, and C/T antigens) could be a population of stem cells that gives rise to primitive tumors (e.g., teratomas, germinal tumors, and pediatric sarcomas). Thus, the fact that VSELs are present in postnatal adult tissues may lend support to the embryonic rest or the germline origin hypotheses of cancer development and reconcile them with currently proposed stem cell theories of cancer development. In fact, markers present in VSELs, such as Oct-4, SSEA, or C/T antigens, have been identified in several malignancies present in pediatric and adult patient tumors. We envision that VSELs can initiate cancer by acquiring mutations, or fusing with other somatic cells. They also can provide stroma and vessels for expanding malingnacies. However, the potential involvement of VSELs in tumorogenesis requires further studies and more direct evidence.